Search

PURPOSE This phase I study (RAD1901-005; NCT02338349) evaluated elacestrant, an investigational oral selective estrogen receptor degrader (SERD), in heavily pretreated women with estrogen receptor–positive, human epidermal growth factor receptor 2–negative metastatic breast cancer, including those with estrogen receptor gene alpha ( ESR1) mutation. The primary objective was to determine the maximum tolerated dose and/or recommended phase II dose (RP2D). METHODS The study consisted of a 3 + 3 design (elacestrant capsules) followed by expansion at RP2D (400-mg capsules, then 400-mg tablets) for the evaluation of safety and antitumor activity. Elacestrant was taken once daily until progression or intolerability. RESULTS Of 57 postmenopausal women enrolled, 50 received RP2D (400 mg once daily): median age, 63 years; median three prior anticancer therapies, including cyclin-dependent kinase 4,6 inhibitors (CDK4/6i; 52%), SERD (52%), and ESR1 mutation (circulating tumor DNA; 50%). No dose-limiting toxicities occurred; the most common adverse events at RP2D (400-mg tablet; n = 24) were nausea (33.3%) and increased blood triglycerides and decreased blood phosphorus (25.0% each). Most adverse events were grade 1-2 in severity. The objective response rate was 19.4% (n = 31 evaluable patients receiving RP2D), 15.0% in patients with prior SERD, 16.7% in patients with prior CDK4/6i, and 33.3% in patients with ESR1 mutation (n = 5/15). The clinical benefit rate (24-week) was 42.6% overall (n = 47 patients receiving RP2D), 56.5% (n = 23, ESR1 mutation), and 30.4% (n = 23, prior CDK4/6i). Elacestrant clinical benefit was associated with decline in ESR1 mutant allele fraction. CONCLUSION Elacestrant 400 mg orally once daily has an acceptable safety profile and demonstrated single-agent activity with confirmed partial responses in heavily pretreated patients with estrogen receptor–positive metastatic breast cancer. Notably, responses were observed in patients with ESR1 mutation as well as those with prior CDK4/6i and prior SERD. A phase III trial investigating elacestrant versus standard endocrine therapy is ongoing.

109 Background: gpNMB is an internalizable transmembrane glycoprotein expressed in melanoma and multiple other tumor types. The ADC GV (CDX-011) delivers the potent cytotoxin MMAE to gpNMB+ cells. GV has shown promising activity in advanced melanoma and breast cancer. Methods: This Phase II study (CDX011-05) assessed the efficacy and safety of GV monotherapy (1.9 mg/kg q3w) for patients (pts) with advanced melanoma progressive after ≤1 chemotherapy, ≥1 checkpoint inhibitor (CPI) and if BRAFV600mutated, ≥1 BRAF/MEK inhibitor. Central IHC determined gpNMB expression in archival and/or pre-treatment tumor. Primary endpoint was objective response rate (ORR) (RECIST 1.1) with ≥6 responders out of 52 evaluable pts as threshold for antitumor activity. Additional endpoints: progression free survival (PFS), overall survival (OS), duration of response (DOR), safety, PK/PD and correlation of tumor gpNMB expression with efficacy. Results: 62 pts enrolled (all evaluable) had median age of 67 years; 55% male; 21% BRAFV600mutated; 63% with ≥3 lines prior therapy; 100% had prior CPI; 100% Stage IV; 89% M1c. One confirmed complete response (CR) and 6 confirmed partial responses (PR, including 1 unconfirmed CR) were seen (confirmed ORR = 11%, p = 0.035 comparing to reference ORR 5%); 33 pts had stable disease including 3 unconfirmed PR. Median DOR = 6.0 (range: 4.1, 14+) months (mos), median PFS = 4.3 mos and median OS = 9.8 mos; 26 pts continue to be followed for survival. All pts with available tissue (60/60) had gpNMB+ tumors; 47/60 had 100% gpNMB+ epithelial cells; no clear correlation with outcome was seen in this population with consistent high expression. Toxicities included alopecia, neuropathy, rash, fatigue and neutropenia. Treatment-related rash in cycle 1 was associated with improved ORR (rash = 22%; no rash = 7%), PFS (p = 0.007) and OS (p = 0.035). Conclusions: GV has promising activity (primary endpoint of ORR was met) with a manageable safety profile in heavily pre-treated melanoma pts. Additional cohorts evaluating GV with either varlilumab, an activating anti-CD27 monoclonal antibody, or PD-1 inhibitors are open to accrual to provide further insights into the synergy of ADC and immunotherapy. Clinical trial information: NCT02302339.

34 Background: TIM-1 (KIM-1, HAVcr-1) is a type 1 transmembrane glycoprotein. Expression is limited in healthy tissues but upregulated after kidney injury, in ovarian cancer and renal cell carcinoma (RCC). TIM-1 is associated with a more malignant RCC phenotype and immune response regulation. CDX-014, an antibody drug conjugate, contains a fully human IgG1 monoclonal antibody (mAb) conjugated to the potent cytotoxin MMAE. CDX-014 is in a Ph 1/2 trial for RCC (NCT02837991). The prevalence of TIM-1 in other cancers was investigated to explore the broader potential for CDX-014. Methods: Tumors and normal adjacent tissues (NAT) were provided (Mosaic Labs) via IRB-reviewed protocol MOS001 for in vitro analysis of remnant, anonymized human samples. The study included 188 tumors: 30 renal, 45 ovarian, 34 endometrial, 2 cervical, 2 vaginal, 15 bile duct, 15 CRC, 30 NSCLC and 15 thymoma plus 75 NAT. Mosaic performed a validated immunohistochemistry assay on FFPE tissues. A FITC-conjugated, anti-TIM-1 mAb was applied after slide deparaffinization and antigen retrieval. Anti-FITC (Novus Bio), Envision+ HRP and DAB (Dako) were used for visualization. A pathologist scored hematoxylin-counterstained slides for % positive cells and staining intensity. A cut-off of > 2% of cells staining at any intensity was deemed positive. Results: TIM-1 was universally prevalent in renal papillary (14/14) and clear cell carcinoma (CCC [13/13]) plus vaginal CCC (2/2). TIM-1 was also seen in ovarian CCC (12/15; 80%), endometrial CCC (9/21; 43%), adeno NSCLC (2/15; 13%), serous ovarian (1/10; 10%), adeno endometrial (1/13; 8%) and CRC (1/15; 7%). There was no staining in chromophobe RCC (3/3), cervical CCC, endometrioid or mucinous ovarian (10 each), bile duct carcinoma, squamous NSCLC or thymoma. There was minimal to no TIM-1 in NAT except for 7/10 kidney samples (possibly from unappreciated kidney injury). Conclusions: The data suggest that in addition to renal and ovarian cancer, TIM-1 targeted therapy such as CDX-014 warrants exploration in patients with endometrial CCC, vaginal CCC and adeno NCSLC. These results support the ongoing first in human Ph 1/2 study in RCC and potential broader clinical development of CDX-014 in other indications.

129 Background: Glycoprotein NMB (gpNMB) is an internalizable transmembrane protein overexpressed in ~20% of BC. gpNMB promoted BC metastases in a murine model and is a poor prognostic marker in BC patients (pts) (Rose 2010). Glembatumumab vedotin (GV, CDX-011) is a novel antibody-drug conjugate targeting gpNMB+ cancer cells with the potent cytotoxin monomethyl auristatin E. In a Phase I/II and in the Phase II EMERGE study, GV was well-tolerated (treatment-related toxicity included rash, neutropenia, and neuropathy) with promising activity in gpNMB+ BC tumors including TNBC. In EMERGE, GV vs. investigator’s choice chemotherapy, demonstrated an objective response rate of 30% (7/23) vs. 9% (1/11) in pts with tumor gpNMB overexpression (gpNMB in > 25% of tumor cells); 18% (5/28) vs. 0% (0/11) in TNBC; and 40% (4/10) vs. 0% (0/6) in gpNMB-overexpressing TNBC. Improvements in progression-free survival (hazard ratio (HR) = 0.11) and overall survival (HR = 0.14) in gpNMB-overexpressing TNBC were noted. Methods: This retrospective analysis of the EMERGE study examined frequency of gpNMB overexpression by various baseline and disease characteristics. Tumors were centrally assessed by immunohistochemistry (IHC) and included data for 328 pts. Results: Tumor gpNMB overexpression was present in 21% (69/328) of screened and 40% (38/96) of TNBC pts. gpNMB overexpression was consistent in progesterone, estrogen, or HER2+ expressing tumors at rates of 12-15%. gpNMB overexpression was not observed among 17 lobular tumors, but was present in 21% (57/270) of ductal tumors. gpNMB overexpression was seen across organ sites, including lung (43%, 3/7), lymph node (31%, 12/39), chest wall (23%, 3/13), breast (21%, 44/209), bone (13%, 1/8), and liver (13%, 3/24). There were no apparent differences in gpNMB overexpression by age, race or disease setting (early vs. advanced) at tissue collection. Analysis is ongoing to determine if prior treatments may upregulate gpNMB expression. Conclusions: gpNMB overexpression in BC, especially in TNBC, appears consistent in archival primary and/or metastatic BC, regardless of age. A randomized multicenter study of GV in gpNMB overexpressing metastatic TNBC (METRIC) is ongoing. Clinical trial information: NCT01156753.

20520 Background: Tumors grow more slowly in endosialin/TEM1 KO mice compared to WT, suggesting that host endosialin/TEM1-positive stroma promotes malignancy (Nanda, PNAS, 2006). Methods: We surveyed 92 human clinical samples by IHC (52 epithelial tumors, 40 sarcomas) to measure the extent of endosialin/TEM1 expression and assess cellular localization. Expression was also studied in primary cell cultures and 49 cancer lines. Results: 46/52 adenocarcinomas and carcinomas (88.5%) expressed endosialin/TEM1 mainly in pericytes (7/7 bladder, 7/11 breast, 5/5 colon, 7/8 ovarian, 7/7 renal, 2/2 liver, 4/5 lung, and 7/7 gastric). In some tumors, endosialin/TEM1 was present in pericytes and in fibroblasts (2/7 bladder, 4/11 breast, 5/5 colon, 5/8 ovarian, 2/7 renal, 2/2 liver, 1/5 lung and 7/7 gastric cancers). 32/40 sarcomas (80%) expressed endosialin/TEM1 mainly in pericytes (10/11 synovial sarcomas, 9/10 malignant fibrous histiocytomas, 3/5 Ewing's sarcomas, 3/5 angiosarcomas, 3/4 rhabdomyosarcomas, 2/2 pleiomorphic sarcomas, 1/1 Kaposi's sarcoma, 0/1 leiomyosarcoma and 1/1 liposarcoma). Synovial sarcomas displayed intense fibroblast staining. Endosialin/TEM1 was expressed by sarcoma cells in a few samples. In culture, we previously showed expression in human fibroblasts, endothelial precursor cells, pericytes and mesenchymal stem cells (MSC). We now report that endosialin/TEM1 mRNA is present in MSC during osteogenic and adipogenic differentiation. In 49 human tumor lines, expression was restricted to sarcoma cells (7/15 sarcoma, 0/5 melanoma, 0/2 neuroblastoma, 0/4 breast, 0/1 ovarian, 0/3 prostate, 0/1 renal, 0/1 liver, 0/3 pancreatic, 0/6 colorectal and 0/8 hematopoietic cancer lines). Conclusions: Endosialin/TEM1 is frequently expressed in clinical cancer (85% of 92 samples), both in epithelial and mesenchymal tumors. Cellular expression is mesenchymal in clinical samples and in culture. In tissues, endosialin/TEM1 is present in pericytes and fibroblasts, with some expression in sarcoma cells. The importance of endosialin/TEM1-positive stroma in experimental animals and prevalence in patient samples warrant further investigation. Endosialin/TEM1 may be a useful therapeutic target. [Table: see text]

13080 Background: Soft tissue sarcomas are malignant tumors that can develop from a variety of tissues. There is a lack of targeted therapies directed specifically against sarcomas with surgery, radiation, and chemotherapy being the standard of care. By employing multiple strategies, new targets that may advance sarcoma therapy can be identified. Methods: Profiling of secreted cytokines that may drive tumor growth was performed on cell lines representing osteosarcomas, liposarcomas, fibrosarcomas, and rhabdomysarcomas. Of 27 cytokines analyzed from cultured media, the following were most highly expressed: IL-6, IL-8, IL-12p40, SDF-1, and MCP-1. Results: Results were compared against gene expression data from publicly available oligonucleotide microarrays derived from human tumors. Ewing’s sarcoma and rhabdomysarcoma samples confirmed higher expression levels of VEGF. Further analysis of array data from 16 classes of sarcomas and 181 tumors indicated that IL-7 was more predominantly expressed than IL-6 and IL-8. To evaluate efficacy of novel agents, xenograft models were established using human cell lines A673 rhabdomyosarcoma and SW872 liposarcoma. Tumor growth and response to therapy was characterized. Paclitaxel (75 mg/kg, days 7, 9, 11) delayed tumor growth for about one week following treatment in both models, however only SW872 liposarcoma tumors responded after dosing with anti-angiogenic agent, Avastin (5 mg/kg, days 4, 7, 11, 14, 17, and 21). Conclusions: In addition to killing cancer cells, destabilization of the vasculature/stroma may be beneficial. Immunohistochemical (IHC) analysis was performed on sarcoma tumors from the clinic. CD31 and SMA staining identified vasculature and stromal regions. Subsequent to identification of novel targets specific to sarcomas, IHC methods can further define localization of antigens between various tumor components. A variety of techniques are being utilized in order to recognize new targets in sarcomas. Cell-based assays and preclinical models are incorporated into the drug development pathway to evaluate novel drugs. Analysis of differential gene expression between tumor types and amongst sarcoma subtypes will be essential in identifying which patient population will gain the greatest benefits from targeted therapies. [Table: see text]